EV revolution powers chip rethink
For Nvidia, it was really not about ICE versus EVs. Until the latter offered OEMs an opportunity to rewrite the rules
A car driver does not expect to have to wait until functions boot up. Or accept sluggishness in its electronics’ performance based on heat or cold.
And that was an initial challenge for Californian chipmaker Nvidia as it moved from consumer electronics into the automotive sector, the firm’s vice-president, automotive Danny Shapiro tells EV inFocus.
“You leave your phone in the car during the summer, you come back and you cannot use it until it cools down,” Shapiro says. “We needed to make sure that the car starts and everything lights up in the car regardless of the temperature in summer or winter.”
More of this direct to your inbox?
Get our free weekly newsletter, plus premium data and content
No spam. Unsubscribe anytime.
But, as Nvidia tackled the challenges of entry into the sector, a new opportunity presented itself. The e-mobility revolution meant both new EV pure plays starting the design process from scratch, and traditional OEMs also grasping the opportunity to rethink the status quo. Shapiro talks us through it.
How has the shift towards EVs in the automotive industry impacted Nvidia?
Shapiro: Automotive is a relatively small part of the company, but it is growing. It is definitely important, and it is something we have been investing in for well over a decade, maybe close to two.
It really had nothing to do with EVs [initially]. We started looking at bringing infotainment systems to market, so advanced graphics. Cars used to just have a radio and dials and buttons, and then smartphones and tablets came along!
We accelerated that move to bring a small computer into the car to drive the graphics systems, the digital cluster, infotainment, rear-seat infotainment systems. It was changing the electronic architecture of the car to have these new systems and trying to create systems that ran software applications.
From that, there started to be, in parallel, this move towards EVs. In many cases, you had new companies that were starting from scratch. They did not have legacy issues, and so they started to design the computing architecture first and then build a car around it.
Tesla was an example of that, of course. We had multiple processors inside Tesla, driving the infotainment, driving the cluster, and driving the driver assistance system, all powered by Nvidia.
Did you have to adapt to managing energy differently in an EV?
Shapiro: I think it probably accelerated for us the need to be even more energy efficient. Any energy we are consuming to run these computer systems takes away range. And so that was really something that was key for us, to make sure that we are as energy efficient as possible.
Were there any challenges unique to EVs that were new to Nvidia when entering this industry?
Shapiro: It was not so much EV versus ICE, but initially we were not an automotive grade company. The entire process — from designing the chips to manufacturing the chips to testing and validation — is a very different process to get to the strict requirements.
For EVs, it is a simpler architecture for the armature; it has just a lot fewer parts. But also it was an opportunity for some companies to say, ‘OK, clean slate, we are creating a new division or a new product line.’
It was an opportunity — put all these sensors in, have a much larger software base and centralise everything so that it is not all these different boxes and components, many of which were fixed function, but rather everything is centralised. And, to update the whole car, you just have to update this one box.
How much has EV penetrated the firm’s business?
The bulk of our customers now are EV, and that is just the nature of new car sales. So many of our customers, JLR and others, have gone on record saying it will all be EV, by a certain point in time.
How important do you consider interoperability to the function of a well-programmed EV?
Obviously, it is important, but these things evolve over time. Because our technology is programmable, there is a lot of flexibility. If it is not part of the system right now, that is OK because it is just a software update that maybe enables that new thing in the future. Just having this open flexible platform is what is really key because it lets you do things in the future that you do not even know are going to exist.
There could be a new app just like on your phone. Who knew that Wordle would be a big thing?! So, we are ready for whatever the next big thing is.